PM 2.5 in Germany

library(tidyverse)
library(lubridate)
library(ggplot2)
library(plotly)
library(grid) #for adjust grid scale

#import data from url or file

# Chiang Mai - http://berkeleyearth.lbl.gov/air-quality/maps/cities/Thailand/Chiang_Mai/Chiang_Mai.txt
# Bangkok - http://berkeleyearth.lbl.gov/air-quality/maps/cities/Thailand/Bangkok/Bangkok.txt

url <- 'http://berkeleyearth.lbl.gov/air-quality/maps/cities/Thailand/Chiang_Mai/Chiang_Mai.txt'
df <- read_tsv(url,skip=10, col_names = FALSE)

Parsed with column specification:
cols(
  X1 = col_double(),
  X2 = col_double(),
  X3 = col_double(),
  X4 = col_double(),
  X5 = col_double(),
  X6 = col_double(),
  X7 = col_double()
)

df

colnames(df) <- c('Year','Month','Day','Hour','PM2_5','X6','X7')
df

df <- df %>%
  select(Year:PM2_5)
df

df <- df %>% 
  mutate(date_time = ISOdate(Year,Month,Day,Hour),
         local_dt = date_time+hours(7),
         local_hr = hour(local_dt))
df

—— Summarized Data (Reserved) ——

df %>% 
  mutate(year = year(local_dt),
         month = month(local_dt),
         date = date(local_dt)) %>%
  group_by(date) %>%
  summarize(avg_pm = mean(PM2_5))

—— Visualized Data ——

## entire data

df %>% 
  mutate(year = year(local_dt),
         month = month(local_dt),
         date = date(local_dt)) %>%
  group_by(date) %>%
  summarize(avg_pm = mean(PM2_5)) %>%
  ggplot(aes(date, y = avg_pm , color = as.factor(year(date))))+geom_line()+geom_smooth()

by date for each month

df %>% 
  mutate(year = year(local_dt),
         month = month(local_dt),
         date = date(local_dt)) %>%
  group_by(year,month,date) %>%
  summarize(avg_pm = mean(PM2_5)) %>%
  filter(year == 2019) %>%
  ggplot(aes(date, y = avg_pm , color = as.factor(month)))+geom_line()

By month for each year

df %>% 
  mutate(year = year(local_dt),
         month = month(local_dt),
         date = date(local_dt)) %>%
  group_by(year,month) %>%
  summarize(avg_pm = mean(PM2_5)) %>%
  ggplot(aes(month, y = avg_pm , color = as.factor(year)))+ geom_smooth(se=FALSE) +scale_x_discrete(limits =month.abb)  

## create heat map

df %>% 
  mutate(date = date(local_dt)) %>%
  group_by(date) %>%
  summarize(avg_pm = mean(PM2_5)) %>%
  mutate(day = day(date),
         month = month(date),
         year = year(date)) %>%
  ggplot(aes(day, y=month, fill=avg_pm)) + geom_tile()

## add factor to map, make it display by day

df %>% 
  mutate(date = date(local_dt)) %>%
  group_by(date) %>%
  summarize(avg_pm = mean(PM2_5)) %>%
  mutate(day = as.factor(day(date)),
         month = as.factor(month(date)),
         year = as.factor(year(date))) %>%
  ggplot(aes(day, y=month, fill=avg_pm)) + geom_tile()

## add border and set a color

df %>% 
  mutate(date = date(local_dt)) %>%
  group_by(date) %>%
  summarize(avg_pm = mean(PM2_5)) %>%
  mutate(day = as.factor(day(date)),
         month = as.factor(month(date)),
         year = as.factor(year(date))) %>%
  ggplot(aes(day, y=month, fill=avg_pm)) +
  geom_tile(color="white") +
  scale_fill_gradient(low = "green" , high = 'red')

adjust cell scale

df %>% 
  mutate(date = date(local_dt)) %>%
  group_by(date) %>%
  summarize(avg_pm = mean(PM2_5)) %>%
  mutate(day = as.factor(day(date)),
         month = as.factor(month(date)),
         year = as.factor(year(date))) %>%
  ggplot(aes(day, y=month, fill=avg_pm)) +
  geom_tile(color="white") +
  scale_fill_gradient(low = "green" , high = 'red') +
  coord_equal()

set interval factor for avg pm 2.5

df %>% 
  mutate(date = date(local_dt)) %>%
  group_by(date) %>%
  summarize(avg_pm = mean(PM2_5)) %>%
  mutate(day = as.factor(day(date)),
         month = as.factor(month(date)),
         year = as.factor(year(date))) %>%
  mutate(pm_lv = cut(avg_pm, ## create new column to classify level of smog
         breaks = c(0,12,35.4,55.4,150.4,250.4,350.4),
         labels = c('good','moderate','unheathy for sensitive groups'
                    ,'unheathy','very unheathy','hazardous'))) %>%
  ggplot(aes(day, y=month, fill=pm_lv)) + ## replace avg_pm with pm_level
  geom_tile(color="white") +
  scale_fill_manual(values = c('green','yellow','orange','red','magenta', 'violet'))+
  coord_equal()

separate by year

df %>% 
  filter(year(local_dt) > 2015) %>%
  mutate(date = date(local_dt)) %>%
  group_by(date) %>%
  summarize(avg_pm = mean(PM2_5)) %>%
  mutate(day = as.factor(day(date)),
         month = as.factor(month(date)),
         year = as.factor(year(date))) %>%
  mutate(pm_lv = cut(avg_pm, 
         breaks = c(0,12,35.4,55.4,150.4,250.4,350.4),
         labels = c('good','moderate','unheathy for sensitive groups'
                    ,'unheathy','very unheathy','hazardous'))) %>%
  ggplot(aes(day, y=month, fill=pm_lv)) + 
  geom_tile(color="white") +
  facet_grid(year~.) + ## separate by year
  scale_fill_manual(values = c('green','yellow','orange','red','magenta', 'violet'))+
  coord_equal()

reverse month -> make less value on top

df %>% 
  filter(year(local_dt) > 2015) %>%
  mutate(date = date(local_dt)) %>%
  group_by(date) %>%
  summarize(avg_pm = mean(PM2_5)) %>%
  mutate(day = as.factor(day(date)),
         month = fct_rev(as_factor(month(date))),  ## remove as.factor and add fct_rev() to month and make it factor using as_factor
         year = as.factor(year(date))) %>%
  mutate(pm_lv = cut(avg_pm, 
         breaks = c(0,12,35.4,55.4,150.4,250.4,350.4),
         labels = c('good','moderate','unheathy for sensitive groups'
                    ,'unheathy','very unheathy','hazardous'))) %>%
  ggplot(aes(day, y=month, fill=pm_lv)) + 
  xlab("Day") + ylab("Month") +
  geom_tile(color="white",width = 2, height=2) +
  facet_grid(year~.) + 
  scale_fill_manual(values = c('green','yellow','orange','red','magenta', 'violet'))+
  coord_equal()

add tooltip

tt <- df %>%  ## put ggplot in to tt (tooltip)
  filter(year(local_dt) > 2017) %>%
  mutate(date = date(local_dt)) %>%
  group_by(date) %>%
  summarize(avg_pm = mean(PM2_5)) %>%
  mutate(day = as.factor(day(date)),
         month = fct_rev(as_factor(month(date))),  
         year = as.factor(year(date))) %>%
  mutate(pm_lv = cut(avg_pm, 
         breaks = c(0,12,35.4,55.4,150.4,250.4,350.4),
         labels = c('good','moderate','unheathy for sensitive groups'
                    ,'unheathy','very unheathy','hazardous'))) %>%
  ggplot(aes(day, y=month, fill=pm_lv)) + 
  xlab("Day") + ylab("Month") +
  geom_tile(color="white") +
  facet_grid(year~.) + 
  scale_fill_manual(values = c('green','yellow','orange','red','magenta', 'violet'))+
  coord_equal()

ggplotly(tt) # display ggplot with tooltip

figure out scale error by increase number of column and set y’s scale to free

tt<-df %>% 
  filter(Year > 2018) %>%
  mutate(date = date(local_dt)) %>%
  group_by(date) %>%
  summarize(avg_pm = mean(PM2_5)) %>%
  mutate(day = as.factor(day(date)),
         month = fct_rev(as_factor(month(date))),
         year = as.factor(year(date)),
         pm_lv = cut(avg_pm, 
           breaks = c(0,12,35.4,55.4,150.4,250.4,350.4),
           labels = c('good','moderate','unheathy for sensitive groups'
                      ,'unheathy','very unheathy','hazardous'))
         ) %>%
  ggplot(aes(day, y=month, fill=pm_lv)) + 
  xlab("Day") + ylab("Month") +
  geom_tile(color="white") +
  facet_wrap(year~., scales = "free", ncol=1) + 
  scale_fill_manual(values = c('green','yellow','orange','red','magenta', 'violet'))

  ggplotly(tt)

like above but force square coord_equal()

tt<-df %>% 
  filter(Year > 2018) %>%
  mutate(date = date(local_dt)) %>%
  group_by(date) %>%
  summarize(avg_pm = mean(PM2_5)) %>%
  mutate(day = as.factor(day(date)),
         month = fct_rev(as_factor(month(date))),
         year = as.factor(year(date)),
         pm_lv = cut(avg_pm, 
           breaks = c(0,12,35.4,55.4,150.4,250.4,350.4),
           labels = c('good','moderate','unheathy for sensitive groups'
                      ,'unheathy','very unheathy','hazardous'))
         ) %>%
  ggplot(aes(day, y=month, fill=pm_lv)) + 
  xlab("Day") + ylab("Month") +
  geom_tile(color="white") +
  facet_wrap(year~., scales = "free", ncol=1) + 
  scale_fill_manual(values = c('green','yellow','orange','red','magenta', 'violet'))+
  coord_equal()

  ggplotly(tt)